BACKGROUND. As for stroke, geographic and ethnic variations in CHD mortality are recognized but poorly understood. NIH (NINDS) funded the Reasons for Geographic And Racial Differences in Stroke (REGARDS) study to follow 30,000 individuals(age >44; 1/2 African American, 1/2 European American; 1/2 female; 1/2 from the stroke belt/ buckle areas) over 3.5 years for stroke incidence and mortality. SPECIFIC AIMS. The overall goal of the REGARDS-MI Ancillary Study is to capitalize on a narrow window of opportunity to leverage REGARD'S infrastructure by quantifying geographic and ethnic variations in CHD mortality, attack rates (new and recurrent hospitalized acute CHD and pre-hospital CHD death in the population) and case fatality (among those with an acute CHD event, death within 28 days). Our Specific Aims are to (1) estimate US geographic- and ethnic-specificrates of CHD mortality, attack rates, and case fatality; (2) study the relationshipbetween US geographic- and ethnic-specific attack rates and variation in the prevalence of CHD risk factors; and (3) study the relationship between US geographic- andethnic-specific CHD case fatality and socioeconomic factors, medical care and CHD risk factors. METHODS. We will retrieve medical records for all heart-related hospitalizations(-6000) among REGARDS participants and collect data on all deaths (-500). We will adjudicatehospitalizations and deaths for CHD, collect cardiac procedures and medications, and merge this with extensive data already being collected by REGARDS. Using multivariable multi-levelanalyses, we will test 10 specific hypotheses. IMPLICATIONS. This will be the first nationwidestudy to relate CHD attack rate and case fatality, including pre-hospital deaths, to geographic and ethnic variations in CHD mortality. REGARDS offers aunique opportunity to attain our specific aims without the massive participant recruitment and follow-up costs that otherwise would be necessary. There is no budgetary overlap between REGARDS and REGARDS-MI as REGARDS focuses only on potential strokes. Our results will contribute significantlyto the knowledge base essential for attaining the Healthy People 2010 goal of reducing health disparities.